Purpose: Despite the prevalent use of the golden ratio in the environment design for autistic users, there is no experimental research that explains their desired space ratio. This study aims to find the desired ratio in the learning environment, like an occupational therapy room for autistic users, while considering effective variables on their choices. Methodology: Based on McAllister's (2012) study, a scale model inspired by a dollhouse has been chosen to work and play with autistic children directly and find their (...) desired space ratio. Meanwhile, the furniture configuration and the access type in the environment could also define the ratio for interviewees. Accordingly, the scale model was designed with six different scale rooms containing three ratios and two spatial access types represented by the furniture arrangement. 50 autistic children of different ages, genders, and disorders in the moderate and medium spectrum have been randomly sampled. Then, they were asked to choose their desired ratio by playing with the scale model and putting the doll in their preferred room. This study also included 50 non-autistic children as a control group with different genders and defined ages. Result: After interviewing both groups, the chi-square test with Cramer's V calculation analyzed the result to find the statistically significant correlation between control variables, among which two independent variables had a statistically significant correlation with the chosen ratio for all the interviewees. First, being autistic significantly impacted their desired ratio choice as autistic children went for a ratio of 2:5, and no autistic ones had a high tendency toward the golden ratio (1:1.6). The chi-square test also showed that the access type in the scale model was the effective variable for both groups when choosing their desired ratio. Conclusion: The access type of the room with its furniture configuration impacted how users perceived the room's ratio, as well as being autistic or non-autistic users. (shrink)